Power-Hardware-in-the-Loop Electric Motor Emulation for Transportation Electrification Considering Mass Production Variations

This paper presents a compact and scalable modeling framework for real-time emulation of permanent magnet synchronous machines (PMSMs) using artificial neural networks (ANNs). The proposed method replaces conventional lookup-table (LUT)-based maps with an ANN to represent the nonlinear flux-current and torque-current relationships of PMSMs. Trained on finite element analysis (FEA)-derived data and fine-tuned using parameter distributions, the proposed framework captures manufacturing variability without requiring extensive retraining or LUT regeneration. In this research, parameter deviation denotes statistical dispersion within production limits and temperature-driven drift; it is not intended to identify or reproduce a specific localized motor defect. The ANN architecture is designed for deployment on FPGA-based real-time simulators, consuming less than 1 kB of memory and 173 floating-point operations (FLOP) per iteration. Experimental validation using a power-hardware-in-the-loop (PHIL) testbed demonstrates excellent agreement between the emulator and measured responses, achieving average mean squared errors below 0.03% across dynamic conditions. This approach offers significant improvements in adaptability, resource efficiency, and scalability over traditional LUT-based models, making it especially well-suited for production-line testing of intelligent drive systems in industrial and automotive applications.